Like many other journalists, I've found "underclass" a handy word. It captures, first, the split in the black community between those who are "making it" and those who are being left behind in the ghettos. Second, it distinguishes the working poor, who have jobs but simply don't make enough money, from those poor people (the underclass) who have a "weak attachment to the labor force," as sociologist William Julius Wilson puts it. "Underclass" is also useful, frankly, because it's respectable, offering blacks, whites, conservatives and liberals a noninflammatory way to note these distinctions.

It may not be respectable for long. In this paper {op-ed, Sept. 10} and elsewhere, the eminent sociologist Herbert Gans has mounted a crusade against the term. "Underclass," he says, "has taken on so many connotations of undeservingness and blameworthiness that it has become hopelessly polluted in meaning, ideological overtone and implications." Gans's criticism recently prompted Wilson to back off and try out an alternative term, "ghetto poor."

Let's concede that "underclass" isn't a perfect word. To my ear, as well as Gans's, it sounds too ominous -- suggesting that once people are caught in this "under" group, they can never escape. That's obviously untrue. As Gans also notes, "underclass" is sometimes used sloppily. Third-generation welfare mothers are part of the underclass. Homeless men, however, have a different set of problems and shouldn't be given the same label.

But Gans has a bigger agenda than finding a more precise word. He is enforcing, through semantic guilt-tripping, a liberal orthodoxy in which distinguishing good behavior from bad among the poor is somehow forbidden. His goal isn't to better describe reality; it's to avoid reality.

Welfare mothers, Gans says, shouldn't be labeled "underclass" because "a welfare recipient often requires little more than a decent paying job . . ." (emphasis added). But welfare mothers who are ready to take jobs, when offered, aren't what people mean by underclass. "Often" an offer of a decent job isn't enough. "Often" a welfare recipient doesn't have the discipline to stick with the job or the right attitude, or she simply prefers to stay on welfare rather than work.

What are we to call people for whom a job isn't enough? Wouldn't it be useful to have a term? Gans may not like the "ideological overtone and implications" of the question, but that doesn't mean these people don't exist. They may not constitute a majority of the poor or even of welfare recipients. But they are a substantial group. Indeed, as more states attempt to nudge welfare mothers into jobs, we hear more and more cries (from liberals, mainly) about how you can't expect these people to work, they need training, lack "interpersonal skills," have "low self-esteem." Right. They do. If they aren't in an "underclass," they're in some kind of class that's in a lot of trouble.

"Underclass," especially as used by Wilson, also implies that the various forms of impoverishing behavior -- out-of-wedlock births, school inattendance, dismissing a $5-an-hour job as "chump change" -- tend to reinforce each other in the social isolation of a ghetto. Does Gans think it makes no difference that the children of single teenage mothers go to school mainly with the children of other single mothers? If Wilson's "isolation" thesis isn't true, Gans should disprove it rather than trying to suppress it.

Gans gave his game away in an academic article he published this summer. Even if "future research should show that some of the very poor are so isolated" that they could "be called an underclass in an analytic sense," he argued, the term would still not be "useful." Why? Because it would imply that other policies were needed "in addition to policies of full employment." In other words, even if the term describes reality, we shouldn't use it because Gans doesn't approve of the policies that might come from it.

What would those policies be? Gans seems to feel they will inevitably be cruel. "Labeling poor people as underclass is to blame them for their poverty," he asserts. In reality, the term has helped produce a growing left-right consensus in favor of replacing welfare with work, supplemented by generous tax credits and subsidized day care. Gans may not like this consensus, but it's hardly abandoning the poor. Yes, it would condemn some underclass behavior, prompting Gans to protest that policy makers "are not hired to be moral judges." But should they really treat as a matter of indifference whether ghetto teenagers stay in school and look for work or instead go on welfare or -- no "moral judging," now! -- abandon their children?

Maybe "underclass" will be replaced by a more precise term. If so, that term will quickly acquire many of the connotations Gans finds unacceptable. "Ghetto poor," for example, puts even more emphasis on the race factor than "underclass." I like the term "welfare culture," because it emphasizes what I think is the key to the problem, but that would send Gans up the wall. Linguistic enforcers like Gans may make us all switch words every few years. But they won't change what people think, until we somehow change the terrible reality of the underclass.

The writer is a senior editor of The New Republic.